Effects of boron concentration on the microstructure, mechanical and tribological properties of powder-pack borided AISI 4140 steel

Three different boron amounts were used in this study to control the microstructure, mechanical, and tribological properties of borided cases on AISI 4140 steel. A single Fe 2 B phase was observed for the lowest boron amount, while both Fe 2 B and FeB phases were present on the other two boron compositions. The highest surface hardness was observed in the highest boron amount. The scratch tests revealed different cohesive failure mechanisms on the samples, such as arc and chevron tensile cracks, and even chipping was observed for all three-studied boron amounts; it appeared at lower applied loads for the highest boron amount, indicating a likely detachment of the fragile FeB layer. The borided layers presented abrasive wear during the ball-on-at wear tests. Despite the lower specic wear rate, the surface damage was more severe in the sample with the highest boron amount, due to its brittleness, resulting in lower resistance to surface fatigue.


Introduction
Wear is one of the most common metal surface degradation mechanisms; there are several ways to promote wear mitigation, such as increasing the surface hardness. Boriding is a very effective surface hardening treatment, which is applied to various ferrous and non-ferrous materials [1]. The boriding process in steels involves heating the material in a range between 700 and 1000°C, for 1 to 12 hours, in contact with a boriding medium such as a solid powder, slurry, liquid, or gas [1]. Boriding in steels generates layers with excellent adhesion [2], mainly due to its columnar morphology [3]. Besides, borided layers present high hardness and, consequently, good wear resistance [4][5][6][7][8][9][10][11][12][13]. One of the boriding challenges is to set the appropriate process parameters to obtain the proper microstructure to withstand a given wear environment.
One highly used commercial powder mixture consists of B 4 C, KBF 4 , and SiC [1]. The mechanism of pack boriding of steels using B 4 C (boron source), KBF 4 (activator), and SiC (diluent) occurs according to the following reactions [37]: The B 4 C and SiC were supplied by Fiven Brazil, while KBF 4 was acquired from Sigma Aldrich. The reactants of each boriding mixture, with a granulometry below 100 µm, were mixed in a Y-mixer for 2 hours at 31.5 rpm on a lathe; and then mixed with 25 mm diameter alumina balls for 1 hour. Subsequently, a portion of each powder mixture was placed in a steel container, enough to ll the container's bottom and create a 25 mm thick powder layer. Afterward, four-steel disc samples were placed in the container, then the rest of the container was lled with the powder mixture, after which the steel container was closed and placed in a furnace at 900°C for 2 h. After the boriding process, the samples were air-cooled.
To facilitate the understanding of the manuscript, the AISI 4140 steel samples were identi ed according to the respective amount of boron in the powder mixture used in the pack boriding, as follows: 1.21B, 4.35B, and 8.26B.

Characterization
The borided specimens were prepared with 320 to 1200 grit SiC abrasive paper and polished with 0. The identi cation of the phases in the borided layers was obtained through X-ray diffraction, XRD (Shimadzu XRD-7000) with CuKα radiation (λ = 1.54 nm), with thin-lm setting using 2 o incidence angle, a scan speed of 0.5°/min and 0.02° step, with a 2θ scan range from 35° to 100°.
The surface hardness (H) was obtained by Vickers indentation, using a 250 mN load and 15 seconds of indentation time. The reduced Young's moduli (E) of the borided cases were extracted from the loaddisplacement curves using the method proposed by Oliver and Pharr [40], by instrumented indentation using a Berkovich indenter, following ASTM E2546-15 Standard [41], applying 24 steps up to a maximum load of 400 mN.

Scratch adhesion tests
The scratches were done and interpreted based on the ASTM C1624-05 Standard [42]. The scratch grooves were obtained in a Revetest Scratch Tester (Anton-Paar Instruments), with a Rockwell C diamond indenter with a tip radius of 200 µm. The test parameters were 398 N/min loading rate, 6 mm/min speed, progressive load from 1 to 200 N, and a scratch length of 3 mm. The scratch grooves were examined in a Tescan Vega 3 SEM to determine the failure mechanisms and to establish the critical loads (L C ).

Tribology tests
The tribological tests were conducted using a ball-on-at con guration of an  (4) and (6) [39].
The SEM images of the three borided conditions are in Fig. 2 Fig. 2 (c) and (f). Although the FeB phase peaks appeared in the XRD pattern (Fig. 1), the SEM images, in Fig. 2 (b) and (

Adhesion
The scratch tests were performed; consequently, the failure modes were identi ed, and the critical loads were assigned for each boriding condition. SEM images were taken to accurately determine the critical loads to reveal the scratch groove details, as shown in Fig. 3. The critical loads were correlated to the de ned and repeated failure mechanisms, as shown in Table 3. Examining the scratch grooves, in Fig. 3, three failure modes can be identi ed in the borided cases: arc tensile cracks, forward chevron tensile cracks, and chipping. The rst failure to occur is arc tensile cracking, which occurs due to a tensile stress eld generated behind the indenter caused by its penetration [47]. The arc tensile cracks were observed at 37 N, 33 N, and 7 N, for the 1.21B, 4.35B, and 8.26B samples, respectively. Arc tensile cracks correlate with Young's moduli (E) difference between the borided layer and steel substrate ( Table 2). The borided layer of the 8.26B sample showed a much higher surface reduced Young's modulus due to the FeB layer, which results in a smaller capacity to deform under tensile stresses, consequently leading to the lowest critical load for this failure mode.  Figure 4 shows the CoF behavior versus the sliding distance. The CoF behavior of the borided cases shows two stages: running-in and steady-state. The running stage varied among the borided cases; the shortest running-in corresponds to the 1.21B sample, and the longest to the 8.26B sample. The variation in the running-in is due to the difference in hardness among the samples since the running-in occurs when the asperities of the Al 2 O 3 ball and the tested borided surfaces are deformed, and the surfaces are brought together. Then, the 1.21B sample has the lowest surface hardness, which resulted in faster deformation of the surface by the alumina counterpart, therefore the shorter running-in. In contrast, the 8.26B sample, with the higher hardness, exhibited more resistance to the asperity's deformation process, and consequently, a longer running-in is observed. In contrast, the steady-state period showed no difference among the three borided cases, and small uctuations are related to oxide formation and its breakage [50], as well as material removal and formation of debris.

Wear behavior
The optical pro ler images and cross-sectional views of the tracks (Fig. 5) were used to obtain the total volumetric wear (V) and then the speci c wear rate (κ) of the tested samples ( Table 4). The wear tracks did not exceed the thickness of the Fe 2 B (for the 1.21B and 4.35B samples) and FeB (for the 8.26B sample) layers. Since borided layers are useful for tribological applications if k < 10 -6 mm 3 /Nm [21], the samples of the three conditions presented mild wear. The 1.21B sample, with the higher wear rate (k=1.3 x 10 -6 mm 3 /Nm), presented the widest wear track, Fig. 5  and (d), is narrower and has a signi cant material agglomeration, mostly at the sides of the track. The 8.26B sample had the lowest wear rate among the samples (k=0.9 x 10 -6 mm 3 /Nm), and the narrowest wear track, as observed in Fig. 5 (e) and (f), which is expected, since the wear rate is usually related to the hardness of the material [51,52], and sample 8.26B, containing the FeB phase, showed the higher surfaces hardness among the studied conditions. Previous results also showed higher wear abrasion resistance of the FeB phase over the Fe 2 B phase [53]. The main wear mechanisms of the borided specimens are shown in Fig. 6 (a), (c), and (e). The rst wear mechanism is grooving, resulting from the plowing effect due to hard particles, which are removed from the borided surface during the tests, and act as an abrasive surface, as previously observed [54], resulting in micro-abrasion of the borided layers. It can be identi ed as scratches parallel to the sliding direction. This mechanism is more severe on the surface of the 1.21B sample but is also present in the 4.35B sample, as revealed in Fig. 6 (a) and (c).
The 1.21B sample showed higher material removal during the tests, resulting in higher grooving levels on its surface, consequently a higher wear rate. The 8.26B sample has a thicker borided layer than the other two samples (Table 2), which decreases the wear loss, as supported by previous results [54][55][56]. The 8.26B sample surface has a large portion of polished regions visible in smooth areas and occurs due to lower levels of material removal, with no visible scratching [21]. The higher surface hardness improved resistance to the plowing effect, which was also observed previously [54].
The repeated loading and unloading of the surfaces can lead to surface fatigue [21], which results in aking and pitting of the surface, the latter occurring when portions of material are torn out, creating irregular craters on the surface namely pits. Although the 8.26B sample has the lowest wear rate, its very hard surface could not sustain large plastic deformation, leading to larger aking of the surface, followed by pitting, as can be seen in Fig. 6 (e). Tribo lms can be observed at the surface of the samples. These lms result from material agglomeration after debris adhere and cluster due to mechanical contacts at the surface, as previously observed [21,51]. Heavily oxidized tribo lms, identi ed as smearing, can be seen inside and on the sides of the wear tracks, mostly in the 8.26B specimen.
The aspect of the Al 2 O 3 balls after sliding is shown in Fig. 7. The images reveal that the sliding contact resulted in the attening of the Al 2 O 3 balls. The wear scar of the alumina ball against 1.21B samples, Fig.   7 (a), shows plowing lines. It indicates grooving as the main wear mechanism, which occurred due to the three-body abrasive wear on the ball and disk, resulting in higher wear of 1.21B disk (borided specimen) ( Table 4). The wear scar of the alumina ball against the 4.35B sample, Fig. 7 (b), shows less grooving than for 1.21B-ball, which indicates a lower formation of debris and less borided specimen wear (Table  4). In the case of the alumina ball against the 8.26B samples, the wear scar was smooth, with no grooving lines. The higher surface hardness of the 8.26B sample resulted in higher alumina's ball wear.
The EDS analysis of the wear scars of the alumina balls shows material adhesion, Fig. 7 (b), (d), and (f). they indicate the presence of mainly iron-oxides. The iron-oxides found in the pair 8.26B sample-ball ( Fig.  6 (f) and 7 (f)), in higher amount, may have reduced the wear in the borided sample. Because the ironoxides lms can act as a lubricant, reducing the CoF; consequently, mitigating wear [21,51]. On the other hand, it was observed high levels of aking and pitting of its surface, probably due to the high brittleness of the FeB phase, resulting in less surface resistance to fatigue.

Conclusions
The mixtures manufactured in this work can be easily reproduced, and new mixtures can be easily obtained in an industrial environment. Furthermore, by controlling the amount of boron in the mixture, the microstructure resulting from the boriding process can also be manipulated. The scratch test revealed the same failure modes regardless of the boron amount: arc tensile cracks, forward chevron tensile cracks, and chipping. On the other hand, the critical loads are different, revealing that microstructure and, consequently, surface hardness, Young's modulus, thickness affect the critical load. The 8.26 wt% B specimens have the most fragile behavior due to the outer FeB phase.
The sliding contact between the borided cases and alumina balls generated abrasive wear. The average friction coe cients are practically the same for the three borided specimens. Although the running-in stage of the friction coe cient behavior is the longest for the 8.26B specimen, which is related to the highest hardness of this borided condition, the higher the hardness, the more di cult is the breakage of asperities.
The speci c wear rate for the 8.26B sample was the lowest among the three investigated conditions due to its higher surface hardness and a larger amount of iron-oxides formation during the tribological tests. On the other hand, the FeB phase increases the brittleness of the borided case causing several different surface fatigue mechanisms.